VMware vSphere Bitfusion virtualizes hardware accelerators such as graphical processing units (GPUs) to provide a pool of shared, network-accessible resources that support artificial intelligence (AI) and machine learning (ML) workloads. vSphere Bitfusion works with artificial intelligence frameworks such as TensorFlow and PyTorch. You can deploy vSphere Bitfusion within a virtual machine or Docker container for use in data center environments. With vSphere Bitfusion, you can monitor health, utilization, efficiency, and availability of all GPU servers in the network. You can also monitor client consumption of GPUs and assign quotas and time limits.
Want to know what is in the current release of vSphere Bitfusion? Look at the latest VMware vSphere Bitfusion release notes.
Learn About Some of Our vSphere Bitfusion Features
Learn the basic concepts of vSphere Bitfusion, and how it virtualizes GPUs and provides a pool of shared compute resources for use by AI and ML applications.
Learn how to add subsequent vSphere Bitfusion servers to a cluster by using the vSphere Bitfusion plug-in.
Learn how to activate vSphere Bitfusion clients in multiple vCenter Server instances or in a Kubernetes cluster.
You can verify that the installation process of vSphere Bitfusion is successful by testing your deployment.
You can upgrade a vSphere Bitfusion cluster without losing your current cluster configuration and monitoring data.
Learn how to license your vSphere Bitfusion servers after the evaluation period.
Learn how to connect your vSphere Bitfusion server to multiple networks by adding, removing, and modifying network interfaces.
You can add a label to a vSphere Bitfusion client to identify workloads.
Learn how to renew the certificates for all vSphere Bitfusion servers and clients.
Learn how to start and stop vSphere Bitfusion applications, and how to allocate GPUs to run multiple applications on the same GPUs. You can also run AI and ML workloads on a specific set of GPUs or servers.
You can use a Paravirtual RDMA (PVRDMA) adapter to improve the performance of your vSphere Bitfusion deployment. RDMA allows applications direct access to the memory from one computer to the memory of another computer without involving the operating system or CPU.
You can test the MTU frame size of your network in vSphere Bitfusion.
You can configure the time period that vSphere Bitfusion retains your detailed and summary usage data. You can download the data by using metric REST APIs.
You can check the performance, stability, available system resources, and software version of a vSphere Bitfusion server by performing a health check. You can also troubleshoot your vSphere Bitfusion environment by examining log files specific to the vSphere Bitfusion server.
Learn how to monitor vSphere Bitfusion by using the graphical user interface provided by the vSphere Bitfusion plug-in within the vSphere Client. You can view current and historical statistics of GPU allocation and usage, memory usage, network traffic statistics, and other data of your vSphere Bitfusion servers and clients. You can also export and download monitoring data as a .csv file to review and troubleshoot your vSphere Bitfusion environment.
Learn how to monitor the vSphere Bitfusion servers in your cluster by using the Monitoring Plug-ins Package. The package contains more than fifty standard plug-ins for monitoring applications, such as Icinga, Naemon, Nagios, Shinken, Sensu.
Learn how to backup and restore the vSphere Bitfusion database. By backing up the database, you can save a snapshot of the configuration, connectivity, health state, and history of your vSphere Bitfusion cluster data. If there is a failure, you can restore the vSphere Bitfusion database and recover the cluster using the snapshot.
Learn how to install and run AI and ML applications with vSphere Bitfusion, and run benchmarks and test to measure the performance of your vSphere Bitfusion deployment. To use TensorFlow, PyTorch, and YOLO, you also install NVIDIA CUDA and NVIDIA CUDA Deep Neural Network library (cuDNN). CUDA is a computing library developed by NVIDIA that enables general computing on GPUs. cuDNN is a GPU-accelerated library of primitives for use with deep neural networks.
Download vSphere Bitfusion
To begin your deployment, download the vSphere Bitfusion appliance and client software packages.
Explore Our Videos
You can learn about deploying and operating vSphere Bitfusion by reading the documentation, and by watching videos on the VMware vSphere YouTube channel.
Learn More About vSphere Bitfusion
- Learn more about vSphere Bitfusion by visiting vSphere Bitfusion Solutions and the Cloud Platform Tech Zone .
- Learn about TensorFlow, an end-to-end open-source platform for machine learning. TensorFlow makes it easy to create machine learning models for desktop, mobile, web, and cloud environments.
- Learn about PyTorch and YOLO. You can use PyTorch to implement an object detector based on YOLO, which is an object detector that uses features learned by a deep convolutional neural network to detect an object.
- vSphere Bitfusion integrates with CUDA, a parallel computing platform developed by NVIDIA for general computing on GPUs. With CUDA, you can dramatically speed up computing applications by harnessing the power of GPUs. Applications developed with CUDA have been deployed to GPUs in embedded systems, workstations, data centers, and in the cloud.
- Understand how NVIDIA cuDNN, a GPU-accelerated library of primitives for use with deep neural networks, integrates with vSphere Bitfusion to accelerate the GPU performance. This integration allows you to focus on training neural networks and developing software applications rather than spending time on low-level GPU performance tuning.
Use vSphere Bitfusion Documentation
The vSphere Bitfusion documents in HTML reflect the latest update release of each major vSphere Bitfusion version. For example, version 2.5 contains all the updates for 2.5.x releases.
You can create custom documentation collections, containing only the content that meets your specific information needs, using MyLibrary.